Tidy Tuesday

This is my nineth contribution to TidyTuesday, which is ‘a weekly podcast and community activity brought to you by the R4DS Online Learning Community’. Their goal is to help R learners learn in real-world contexts.

For more information, visit the TidyTuesday homepage, check out their GitHub repository and follow the R4DS Learning Community on Twitter.

The purpose of these posts is mainly for exercising purposes. Thus, the provided graphs are not necessarily designed to provide the greatest possible insights. However, I always provide the R code for interested people at the page bottom.

Scooby Doo Episodes

This week’s data comes from Kaggle and was aggregated by plummye. It contains the dates of countries when they declared independence.

Word Cloud for Episode Titles

The dataset is very large and one could spend a lot of time investigating it. I am interested in which words are used frequently in the episodes’ titles. To investigate this descriptively, I plot the frequency of the words in the titles as a word cloud using the wordcloud2-package. To make the word cloud take the shape of scooby, I provide a shape image containing the silhouette of Scooby, which was received here:

# filter out uninteresting words and interpret "scooby's" as "scooby"
scooby$title <- gsub("Scooby's", "scooby", scooby$title) 
scooby$title <- gsub("and", " ", scooby$title) 
scooby$title <- gsub("The", " ", scooby$title) 
scooby$title <- gsub("the", " ", scooby$title) 
scooby$title <- gsub("for", " ", scooby$title) 
scooby$title <- gsub("from", " ", scooby$title) 
# convert to corpus:
docs <- Corpus(VectorSource(scooby$title))
# Convert the text to lower case
docs <- tm_map(docs, content_transformer(tolower))
# Remove numbers
docs <- tm_map(docs, removeNumbers)
# Remove punctuations, but keep apostrophes (e.g. don't, not dont)
docs <- tm_map(docs, removePunctuation, preserve_intra_word_contractions = TRUE)
# Eliminate extra white spaces
docs <- tm_map(docs, stripWhitespace)

# create term document matrix
dtm <- TermDocumentMatrix(docs)
m <- as.matrix(dtm)
v <- sort(rowSums(m),decreasing=TRUE)
d <- data.frame(word = names(v),freq=v)
# reduce frequeny, so that words fit into graphics device (not ideal!)
d$freq[1:2] <- 30
# use different colors
cols <- hcl.colors(982, palette = "Green-Orange")
set.seed(17721)
w <- wordcloud2(d, figPath = "https://github.com/jgranna/tidytuesday/blob/main/2021-07-13/_images/scooby.jpg", size = 1, backgroundColor="black", minSize = 1, color = cols)

Scooby Doo

It is apparent, that “scoobydoo” is used frequently in the episodes’ titles. But also the words “mystery”, “night”, “ghost”, etc. occur quite often.

One could spend much more time to improve the image. Also, the plotting only worked for me in the browser and thus I could not export the image in a “nice way”, but had to export it manually from the browser, which is of course not ideal.

Full R code available on Github.

References